Fast encoding algorithm based on depth of coding-unit for high efficiency video coding
Fast encoding algorithm based on depth of coding-unit for high efficiency video coding
复制标题
DOI：
10.1117/1.oe.51.6.067402
复制
发表时间：
2012-06
影响因子：
1.3
通讯作者：
Han Soo Lee;K. Kim;Tae Ryong Kim;G. Park
中科院分区：
文献类型：
--
作者：
Han Soo Lee;K. Kim;Tae Ryong Kim;G. Park
在高效视频编码（HEVC）方案中，最大编码单元（LCU）的拆分结构由于其递归四叉树表示而高度可变且费力。因此，HEVC编码器的计算复杂度显著增加。提出了一种快速HEVC编码方法，通过使用参考帧或时间上先前帧中的同位LCU的相同结构来预测当前LCU的分裂结构，以降低编码计算复杂度。仿真结果表明，与原HEVC编码方案相比，该方法在平均PSNR损失和比特率增加的情况下，编码时间平均减少了21.3%.
In high efficiency video coding (HEVC) scheme, the split structure of the largest coding-unit (LCU) is highly variable and laborious due to its recursive quad-tree representation. Thus, the computational complexity of the HEVC encoder increases dramatically. A fast HEVC encoding method is proposed to predict the split structure of the current LCU by using the same structure of the co-located LCU in the reference or temporally-previous frame, in order to reduce the encoding computational complexity. Simulation results show that the proposed method can achieve a reduction in encoding time of 21.3% on average, with a negligible average PSNR loss and a negligible bit rate increase compared with the original HEVC encoding scheme.